Annexin A2 is involved in antiphospholipid antibody-mediated pathogenic effects in vitro and in vivo

Abstract

Antiphospholipid (aPL) antibodies recognize receptor-bound beta(2) glycoprotein I (beta(2)GPI) on target cells, and induce an intracellular signaling and a procoagulant/proinflammatory phenotype that leads to thrombosis. Evidence indicates that annexin A2 (A2), a receptor for tissue plasminogen activator and plasminogen, binds beta(2)GPI on target cells. However, whether A2 mediates pathogenic effects of aPL antibodies in vivo is unknown. In this work, we studied the effects of human aPL antibodies in A2-deficient (A2(-/-)) mice. A2(-/-) and A2(+/+) mice were injected with immunoglobulin G (IgG) isolated from either a patient with antiphospholipid syndrome (IgG-APS), a healthy control subject (IgG-normal human serum), a monoclonal anti-beta(2)GPI antibody (4C5), an anti-A2 monoclonal antibody, or monoclonal antibody of irrelevant specificity as control. We found that, after IgG-APS or 4C5 injections and vascular injury, mean thrombus size was significantly smaller and tissue factor activity was significantly less in A2(-/-) mice compared with A2(+/+) mice. The expression of vascular cell adhesion molecule-1 induced by IgG-APS or 4C5 in explanted A2(-/-) aorta was also significantly reduced compared with A2(+/+) mice. Interestingly, anti-A2 monoclonal antibody significantly decreased aPL-induced expression of intercellular cell adhesion molecule-1, E-selectin, and tissue factor activity on cultured endothelial cells. Together, these data indicate for the first time that A2 mediates the pathogenic effects of aPL antibodies in vivo and in vitro APS.

Figure 1

Analysis of antigen specificity of the monoclonal Ab preparations. (A) ELISA analysis of antigen specificity. IgG-APS (500 μg/mL), IgG-NHS (500 μg/mL), 4C5 (100 μg/mL), MuMoAbC (100 μg/mL), and anti-A2 MoAb (100 μg/mL) were diluted 1/50 for aCL and 1/100 for anti-β₂GPI and anti-A2 assays, and tested by ELISA, as described in “Preparation of immunoglobulin G” and “Western blotting to detect antigen specificity.” Results are expressed as means ± SD O.D. units. (B) Western blot analysis of antigen specificity. Proteins (25 μg/lane from HUVECs, A2^+/+, and A2^−/−) as well as 200 ng/lane purified β₂GPI were resolved on 10% sodium dodecyl sulfate–polyacrylamide gels and blotted with the respective IgGs (4C5 and MuMoAbC). To control for loading, the same blot was stripped and probed with monoclonal IgG directed against A2 and glyceraldehyde-3-phosphate dehydrogenase.

Figure 2

Effect of anti-A2 on IgG-APS–induced expression of ICAM-1 and E-sel in HUVECs. HUVECs were cultured and treated with either 200 μg/mL IgG-APS or 200 μg/mL IgG-NHS in the presence or absence of 1 μg/mL anti-A2 IgG Ab or 1 μg/mL MuMoAbC, as indicated in “In vitro detection of surface E-sel and ICAM-1 on EC.” Some cells were treated with LPS, as a positive control, or with medium alone, as a negative control. ICAM-1 (A) and E-sel (B) expression were determined by cyto-ELISA, and the data expressed as means ± SD in O.D. units. Experiments were run in triplicate and performed 3 times. ¶Statistically different from medium-treated cells (P = .001). *Statistically different from IgG-NHS–treated cells (P = .001). **Statistically different from IgG-APS–treated cells (ICAM-1, P = .001; E-sel, P = .001).

Figure 3

Effect of anti-A2 Ab on TF activity induced by IgG-APS in HUVECs. HUVECs were cultured and treated with either IgG-APS or IgG-NHS in the presence or absence of anti-A2 MoAb, as indicated in “TF functional assay.” Some cells were treated with LPS, as a positive control, or with medium alone, as a negative control. TF activity was determined using a chromogenic assay, and data expressed as fold increase (mean ± SD) over the corresponding control. Experiments were performed 3 times in duplicate. ¶Statistically different from medium-treated cells (P = .031). *Statistically different from IgG-NHS–treated cells (P = .015). **Statistically different from IgG-APS–treated cells (P < .044; P < .05).

Figure 4

Effect of aPL Abs on thrombus formation in A2^−/− mice. A2^−/− or A2^+/+ mice were treated with either IgG-APS, IgG-NHS (A), 4C5, anti-A2 MoAb, or MuMoAbC as control (B), as indicated in “Analysis of thrombus dynamics.” Thrombi were induced in the animals, and thrombus size was measured in square microns (μm²). The data are expressed as means ± SD (5-10 animals were used per group). ¶Statistically different from A2^+/+ mice treated with IgG-NHS (A; P = .001) or MuMoAbC (B; P = .002). *Statistically different from A2^+/+ mice treated with IgG-APS (A; P = .002) or 4C5 (B; P = .005). **Statistically different from A2^−/− mice treated with IgG-APS (A; P = .001) or 4C5 (B; P = .009).

Figure 5

Effect of aPL Abs on TF activity in carotid artery homogenates in A2^−/− mice. A2^−/− or A2^+/+ mice were treated with IgG-APS, IgG-NHS (A), 4C5, anti-A2 MoAb, or MuMoAbC as control (B). TF activity was determined in homogenates of carotid arteries using a chromogenic assay, and data expressed as means ± SD in pmol/mg per mL⁻¹ protein. Experiments were assayed in duplicate and repeated thrice. ¶Statistically different from A2^+/+ mice treated with IgG-NHS (A; P = .015) or MuMoAbC (B; P = .001). *Statistically different from A2^+/+ mice treated with IgG-APS (A; P = .026) or 4C5 (B; P = .007). **Statistically different from A2^−/− mice treated with IgG-APS (A; P = .046) or 4C5 (B; P = .002).

Figure 6

Aortic VCAM-1 expression in A2^−/− and A2^+/+ mice treated with IgG-APS. A2^−/− or A2^+/+ mice were treated with either IgG-APS, IgG-NHS, LPS, or PBS, as described in “Determination of endothelial VCAM-1 expression in en face preparations of mouse aoura using quautum dot bioconjugates and α-photon excitation laser-scanning microscopy.” VCAM-1 expression was determined by immunostaining using a specific quantum dot bioconjugate, and quantified after examination using a dual photon laser confocal microscope. Fluorescence intensity in AU (mean ± SD; n = 10 images/mouse and 2 mice/group). ¶Statistically different from A2^+/+ treated with PBS (P = .001). ¶¶Statistically different from A2^−/− treated with PBS (P = .017). *Statistically different from A2^+/+ mice treated with IgG-NHS (A; P = .001) or MuMoAbC (B; P = .001). **Statistically different from A2^+/+ mice treated with IgG-APS (A; P = .001) or 4C5 (B; P = .017).

Figure 7

Representative images of endothelial VCAM-1 expression among the different treatment groups. The red fluorescent staining (655 nm Qdot bioconjugate) indicates surface VCAM-1 immunoreactivity, whereas blue staining represents nuclear Hoechst staining. (A) A2^+/+ or A2^−/− mice were treated with either IgG-APS or IgG-NHS, and LPS or PBS were used as signal controls. (B) A2^+/+ or A2^−/− mice were treated with 4C5, anti-A2 MoAb, or MuMoAbC as control, and LPS or PBS were used as signal controls.

Figure 8

Diagrammatic representation of interaction of aPL/anti-β₂GPI Abs with β₂GPI and receptor(s) on ECs. aPL/anti-β₂GPI Abs bind to domain I (DI) of β₂GPI. β₂GPI anchors to annexin A2 on the surface of ECs, possibly through domain V (DV) of the protein. Annexin A2 does not have a transmembrane domain. Hence, it is not able to transduce intracellular signaling. Other membrane proteins (ie, TLR-4 or apoER2′) may act as accessory molecules or may bind β₂GPI directly and induce phosphorylation of p38 mitogen-activated protein kinase (p38 MAPK) and translocation of nuclear factor-κB (NF-κB), leading to a proinflammatory/prothrombotic phenotype (up-regulation of tissue factor and cellular adhesion molecules [ie, E-sel, ICAM-1, VCAM-1]). Reprinted from Trends in Immunology, P. Meroni, N. Rhonda, E. Raschi, and M.O. Borghi, Humoral immunity against endothelium; theory or reality? 2005;26:275-281, with permission from Elsevier.